The present invention relates to design modifications to prior art or newly designed RFID interrogation systems for protecting a medical device against RFID-associated electromagnetic interference (EMI). More particularly, the novel RFID interrogation systems include a radio frequency identification (REID) communicator which has a circuit for limiting the total continuous transmit time of an electromagnetic signal, and a time-out circuit for delaying a subsequent transmission of the electromagnetic signal.
The RFID reader industry has literally been exploding over the last few years with new applications and indications being discovered on what sometimes almost seems a daily basis. For example, RFID readers and their associated tags are being used for inventory tracking, pharmaceutical tracking, tracking of patients in hospitals, automated checkout in super markets of a basket full of goods with associated RFID tags, automobile keyless entry systems and keyless ignition systems, operating room sponge detector systems, and identification of patient RFID wrist bands. There are several main frequency bands that are now dominating the worldwide RFID industry. Four of the popular ones are low frequency (LF) which generally ranges from 125 to 150 kHz, high frequency (HF) which is at 13.56 MHz, very high frequency (VHF) which is at 433 MHz, and ultra high frequency (UHF) which generally operates at 915 MHz. Moreover, there are both national (American) and international standards (ISO) defining the modulation protocols and pulse widths and repetition rates so that standardized RFID tags can be read by a wide variety of readers. In fact, many readers transmit over a broad range of the RFID protocols for this exact reason. With the explosion of RFID emitters (readers also known as interrogators and sometimes referred to herein as communicators), patients with passive or active (electronic) medical devices (PMDs or AMDs) are increasingly running the risk of coming in close contact with such emitters. AMDs can also be implanted inside (or partially inside) the human body and are known as active implantable medical devices (AIMDs).
FIG. 1 is a wire formed diagram of a generic human body. Various locations are shown for active, passive, structural and other implantable and external medical devices 10 that are currently in use, and in which the present invention may find application. 10A represents a family of external and implantable hearing devices which can include the group of hearing aids, cochlear implants, piezoelectric sound bridge transducers and the like. 10B includes an entire variety of neurostimulators and brain stimulators, and hydrocephalic fluid pumps, drug and hormone insulin injection administration devices, etc. Neurostimulators are used, for example, to stimulate the Vagus nerve to treat epilepsy, obesity, Parkinsonism and depression. Brain stimulator systems are similar to a pacemaker-like pulse generator and include leads leading to electrodes implanted deep into the brain. One application involves sensing of the onset of abnormal SNS electrical activity and then providing electrical stimulation to brain tissue to abort the seizure. The electrodes on the end of the leads that arise from a deep brain stimulator are often positioned in the brain tissue using imaging, most commonly during real time MRI. 10C shows a cardiac pacemaker which is well-known in the art. 10D includes the various types of left ventricular assist devices (LVAD's), and artificial hearts, for example, the recently introduced centrifugal empowered devices. 10E includes an entire family of drug pumps which can be used for dispensing of insulin, chemotherapy drugs, pain medications and the like. Insulin pumps are evolving from passive devices to active or semi-active devices that have sensors and closed loop systems wherein real time monitoring of blood sugar levels is associated with directly related and programmable dose responses. These devices tend to be more sensitive to EMI than passive pumps that have no sense circuitry or transcutaneous leads. 10F includes a variety of external or implantable bone growth stimulators for rapid healing of fractures. 10G includes urinary and/or fecal incontinence devices. 10H includes the family of pain relief spinal cord stimulators and anti-tremor stimulators. 10H also includes an entire family of other types of neurostimulators used to block pain. 10I is representative of implantable cardioverter defibrillators (ICDs) including those with biventricular and multi-site synchronization capabilities for the treatment of congestive heart failure (CHF). 10J illustrates an externally worn device. This external pack could be an insulin or other drug pump, an external neurostimulator or pain suppression device, a Holter monitor with skin electrodes or even a ventricular assist device power pack. 10K illustrates the insertion of transcutaneous probe or catheter. These devices can be inserted into the femoral vein, for example, or into many other endovascular or endothelial lined cavities in the human body.
It would be highly undesirable for any type of AIMD to malfunction when exposed to an RFID reader. This includes externally worn AIMDs such as a Holter monitor. It would be very confusing for medical personnel to interrogate the Holter monitor which stores, for example, cardiac electrograms, and see what they thought was a sustained dangerous cardiac arrhythmia which was in fact due to persistent EMI from an RFID interrogation.
It has been demonstrated that RFID communicators, such as RFID readers, interrogators and emitters, can interfere with medical devices such as implanted cardiac pacemakers and implantable cardioverter defibrillators (ICDs). Initial studies conducted by the inventors have been corroborated through two extensive studies at the FDA. Center for Devices and Radiological Health (FDA-CDRH). In laboratory studies in 2006 and 2008 at the FDA-CDRH, it was determined that RFID readers can and do cause potentially serious EMI to both cardiac pacemakers and ICDs. The FDA report entitled “In Vitro Tests Reveal Sample Radio Frequency Identification. Readers Inducing Clinically Significant Electromagnetic Interference to Implantable Pacemakers and Implantable Cardioverter Defibrillators” is slated to be published in The Heart Rhythm Society journal. The FDA, in its 2008 study, referenced an article published in the New England journal of Medicine on May 27, 1997. This was a seminal paper authored by Dr. David Hayes, et al. where the possible types of responses to EMI of both pacemakers and ICDs were analyzed and classified. The paper classified the EMI responses into a Type 1, Type 2 or Type 3 responses. Type 1 responses were defined as those types of EMI responses that could or would be highly clinically significant including life-threatening responses. Other types of responses, which could simply be annoying, were categorized as Type 2, and others, Type 3, are really of no relevant clinical significance. An example of a Type 3 response would be when a pacemaker detects that EMI is present and goes into a fixed rate safety pacing mode (also known as noise reversion). This is not particularly desirable, but it is also not harmful to the patient for short periods of time. However, a Type 1 response would include, for example, prolonged pacemaker inhibition. This would mean that the pacemaker stopped delivering its life-giving output pulses. This could very quickly be life-threatening for a pacemaker-dependent patient.
Almost all modern pacemakers and ICDs incorporate feedthrough capacitor EMI filters to protect them against high frequency emitters, such as cellular telephones, microwave ovens and the like. U.S. Pat. Nos. 5,333,095; 4,424,551; and 6,765,779 illustrate and describe examples of such prior art feedthrough capacitor EMI filters
FIG. 2 illustrates a prior art unipolar hermetic terminal 20 typically used in active implantable medical devices. Hermetic terminals typically consist of an alumina insulator 22 which is gold brazed 24 to a ferrule 26. In turn, the ferrule is typically laser welded 28 to the titanium housing 30 of an active implantable medical device. There is also a hermetic seal 32 that is formed between the alumina insulator 22 and the lead 34. This is typically also done by gold brazing, glass sealing or the like. There is also a prior art ceramic feedthrough capacitor 36 shown co-bonded to the hermetic terminal subassembly. Such feedthrough capacitors 36 are well known in the art for decoupling and shielding against undesirable electromagnetic interference (EMI) signals, such as those produced by cellular telephones, microwave ovens and the like. See, for example, U.S. Pat. Nos. 4,424,551; 5,333,095; 5,905,627; 6,275,369; 6,566,978 and 6,765,779.
FIG. 3 is a partial cutaway view showing the details of the prior art feedthrough capacitor 36 of FIG. 2. One can see that it has internally embedded electrode plate sets 38 and 40. Electrode plate set 40 is known as the around electrode plate set and is coupled to the capacitor's outside diameter metallization 42. The active electrode plate set 38 is electrically connected to the capacitor inside diameter metallization 44.
FIG. 4 is a schematic diagram of the prior art feedthrough capacitor 36 illustrated in FIGS. 2 and 3. Prior art feedthrough capacitor EMI filters are generally of relatively low capacitance value (generally below 10,000 picofarads). As shown in schematic of FIG. 4, it forms what is known in the art as a single element low pass filter.
Due to size and other limitations, the capacitance value of these prior art low pass feedthrough capacitors is relatively low in value. Because of its low capacitance value, the filter is not effective at attenuating low frequencies, such as for LF readers. In fact, in the LF reader frequency band of 125 to 135 kHz, prior art feedthrough capacitor filters provide less than 0.5 dB of attenuation. These prior art filters are particularly effective, however, for UHF readers operating at 915 MHz. In these bands, the AIMD filter provides well over 30 dB of attenuation and in many cases, above 50 dB.
The results from the FDA studies of pacemakers and ICDs with RFID readers exactly correlate with this. There were no Type 1 responses for any UHF reader operating at 915 MHz. However, for LF and HF readers, the FDA documented a high number of life-threatening Type 1 responses out to a distance of 60 cm.
FIG. 5 is a family of curves which illustrates the performance of the prior art feedthrough capacitors illustrated in FIGS. 2, 3 and 4. In FIG. 5, one can see that the attenuation in decibels (dB) varies with frequency. These are also known in the art as single element low pass filters. In prior art pacemakers and implantable defibrillators, the inventors have found that the value of the feedthrough capacitor, which is intended to provide protection to EMI from cellular telephones, generally varies from 400 picofarads up to about 4400 picofarads (a very few designs go as high as 10,000 picofarads). One can see in FIG. 5 that at 915 MHz, all of the feedthrough capacitor values offer substantial attenuation (above 30 dB). This is why in the FDA studies, no clinically significant Type 1 EMI responses to pacemakers and ICDs at the 915 MHz RFID frequency were found. However, when one examines the 13.56 MHz frequency, one will see that high value feedthrough capacitors in the range of 2700 to 4400 picofarads) offer a substantial amount of attenuation which varies from 17 dB to approximately 23 dB. However, some pacemaker/ICD manufacturers use relatively low value feedthrough filters in the 400 to 1200 picofarad range. In general, those do not offer sufficient attenuation at 13.56 MHz. This is why some manufacturers of pacemakers and ICDs exhibited no problems during the FDA HF testing (no Type 1 responses) whereas other pacemakers did show Type 1 responses. It should also be noted that on FIG. 5, LF (125 to 135 kHz) is substantially to the left (not shown) on the frequency axis. In FIG. 5, the frequency axis starts at 1 MHz and goes up to 915 MHz. For LF, no matter what the value of the feedthrough capacitor (from 400 to 10,000 picofarads) the attenuation is less than 0.5 dB. In other words, prior art feedthrough capacitors are totally ineffective at LF RFID frequencies and there is virtually no passive filter protection at LF frequencies at all for pacemakers and ICDs.
Passive filters include capacitors, inductors and resistors. The word “passive” means that, unlike electronic active filters, passive filters do not require a power source. Passive filters are preferred for EMI low pass filters because they can handle very high amplitude signals (like EMI from cellular phones or RFID readers) without becoming non-linear. Active filters can be designed to operate at LF frequencies. However, since they are based on very low voltage micro electronic circuit chips, they have a very limited dynamic range. It has been demonstrated that active filters become very non-linear and ineffective in the presence of high amplitude signals such as those produced by cellular phones or RFID readers. Accordingly, the AIMD manufacturer really does not have any practical design options to provide effective EMI filtering at LF RFID reader frequencies. Active filters become non-linear in the presence of high intensity RFID fields which rule them out. For an implanted passive filter to be effective at LF, it would need to be several orders of magnitude higher in capacitance value compared to prior art feedthrough capacitor filters. This would make it much too large in both volume and weight (the passive filters would almost be the size of a modern pacemaker). Worse yet, such passive filtering on the therapy delivery or sense circuits of a pacemaker or ICD would degrade its essential performance (pulse degradation, ability to sense biologic signals, etc.).
FIG. 6 illustrates typical (text book) sensing curves for both pacemakers and ICDs. The approximate center of these curves, where the devices are the most sensitive, is around 10 to 100 Hz. This means that signals that fall within this passband are meant to be sensed by the pacemaker. In the case of an ICD, this would be down in amplitude as low as approximately 100 micro-volts and for a pacemaker approximately 0.8 millivolts. This is a range of biologic frequencies that are produced by the human heart. It is important that the pacemaker sense these frequencies so that it can inhibit in the presence of a proper heart beat (proper sinus rhythm). This is an important battery saving feature as there is no reason for the PG to supply voltage pulses if the patient has his own “normal” sinus rhythm. This also prevents a condition called rate competition, which is a situation where if the pacemaker did not inhibit and the pacemaker patient had his own return to sinus rhythm, the pacemaker would actually compete with the patient's underlying rhythm.
It is instructive to look at FIG. 6 and reflect on what happened a number of years ago when there were numerous reports of cellular telephones interfering with cardiac pacemakers and ICDs, Obviously a cellular telephone transmits at much higher frequency than that as illustrated in the sensing curves shown in FIG. 6. However, what happens is that a high frequency carrier, such as that of a cellular telephone which is around 1000 MHz, would enter into previously unfiltered pacemakers and encounter a nonlinear circuit element such as a protection diode. These nonlinear circuit elements act as a demodulator. One of the worst offenders was the old TDMA 11 Hz modulated cellular telephone. Even though it operated at very high frequency, the nonlinear diode elements of a pacemaker would demodulate or strip off the 11 Hz modulation signal, which would fit right into the sensitive portion of the pacemaker passband of FIG. 6 and be oversensed. Oversensing means that the pacemaker would incorrectly interpret this 11 Hz EMI modulation as a normal cardiac heartbeat and inhibit. This is particularly life-threatening for a pacemaker dependent patient whose every heart beat depends on a proper pulse from the pacemaker. Having the pacemaker stop working or inhibit in this situation is immediately life-threatening.
With this understanding, one can now look at the table of FIG. 7. It is extremely unfortunate that the RFID industry has chosen modulation frequencies that fall generally in the range that would fit into the most sensitive portions of both the ICD and pacemaker passband sensing curves. For example, referring to FIG. 7, one sees listed here thirteen different types of RFID readers that were recently tested by the FDA shown in the left hand column. For example, RFID Equipment Code 2 operates at 134 kHz (0.134 MHz), but has a modulation of 14.43 Hz. It was predicted by the members of the Association for the Advancement of Medical Instrumentation Pacemaker Electromagnetic Interference Task Force PC69, that this was likely to be a problem. In fact, in the FDA laboratory tests, all of the LF and many of the HF RFID readers that had pulse repetition rates within the pacemaker passbands indeed caused pacemaker inhibition and/or other types of highly clinical significant Type 1 life-threatening responses. It is also interesting to note, referring to FIG. 7, that the readers that are marked CW (continuous wave) have no modulation content, These CW readers exhibited no Type 1, 2 or 3 responses to pacemakers or ICDs. One might be tempted to immediately lump to the conclusion that a simple way around this entire problem would be to simply restrict the RFID industry to only use CW readers. The problem with that is that CW readers, by definition, can only activate a tag and detect the presence of a tag and can obtain only very limited information. In other words, they can't really transmit back and forth (read/write) any detailed useful information. Accordingly, use of CW tags and readers will not allow for full identification of model number, serial number, and patient information related to an AIMD.
The FDA has conducted two extensive trials testing both pacemakers and ICDs in a laboratory environment wherein cardiac pacemakers and implantable cardioverter defibrillators (ICDs) and their associated leads were placed into human phantom saline tanks and exposed to various model RFID readers and associated systems. FIG. 8 summarizes the testing that was performed by the FDA-CDRH in 2006 and 2008. There were a total of 37 pacemakers and 34 ICDs tested. A total of 20 RFID systems were also evaluated. This testing was blinded in that the results were given letter codes so that no one reading the reports could tell who the manufacturer of the particular pacemaker was or who the manufacturer of the particular RFID system was. Referring once again to FIG. 8, it should be noted that all the major pacemaker and 100 manufacturers in the world participated in this testing by providing their devices.
FIG. 9 is a top down view of a grid placed over the saline tank used for this testing at the FDA. In 2006, a spiral lead configuration was used in accordance with ANSI/AAMI Standard PC69. In 2008, a more representational human implant geometry was used based on the distances of the lead bodies and electrodes from the pulse generator observed on patient X-rays.
FIG. 10 shows a similar set up for ICDs. On the right hand figure (2008), one can see a loop L representative of where excess leadwire would be wound up either in or adjacent to a left pectoral ICD pocket. The configuration is typical for leads passing from the left pectoral region to terminate in the right ventricle and right atrium.
FIG. 11 is a cross-section of the human phantom saline tank showing the implant (pacemaker or ICD) just below (0.5 cm) the surface of the fluid. It has been shown in the past that this type of model very accurately represents the fields that will occur inside the human body. Saline solution of 500 ohm-cm is used in the tank, which replicates the dielectric properties of body fluids. Thus, such a saline tank closely replicates the EMI characteristics encountered by a device that has been implanted inside the human body. The testing, as performed by the FDA, was done with the antenna suspended in a robotic arm that could carefully step the RFID antenna away in discrete distances so that accurate threshold distances for Type 1, 2 and 3 responses could be recorded.
The following definitions are provided to assist with a better understanding of the ratings applied to the FDA electrocardiogram (EKG) tracings in FIGS. 12 through 20. Class 1 responses include transient ventricular inhibition for 3 seconds or more, persistent ventricular inhibition or any change in pulse generator programmed settings. It should be noted that throughout all of the RFID testing, to be discussed in more detail below, there was never a change in programmed settings. In other words, when the RFID reader was removed or turned off, the EMI response immediately ceased. Class 2 response is defined as transient (intermittent) ventricular inhibition for more than 2 seconds, but less than 3 seconds, A Class 2 response also included transient, continuous atrial inhibition or rate adaptive pacing. Class 2 responses are not considered to be immediately life threatening, but are considered to be very undesirable. For example, persistent very high rate stimulation of the heart can produce irreversible damage and/or death, particularly in heart failure patients, patients with ischemic heart disease and others. Class 3 is defined as any other type of interference which includes transient inhibition of less than 2 seconds, and/or noise reversion mode pacing. This is a software circuit feature where the pacemaker detects EMI and reverts to a non-responsive fixed rate (metronome like) stimulation. This is similar to what occurs after a magnet is applied over a pacemaker to close a reed switch, for example, to bypass the circuit that ordinarily decides whether pacing is required or not based on the feedback electrical signals that are being received from the heart. Class 3 responses are undesirable but not considered to be clinically significant.
FIG. 12 shows an EKG base line strip of a pacemaker in the saline tank without an RFID reader present. Normal pacing pulses are shown with the atrial pulses A shown on top and the ventricular pulses V shown on the bottom. One can see that every time there is an atrial pulse, a ventricular pulse follows after the programmed delay (PD). This A-V delay is a normal function as also occurs naturally during sinus rhythm within the human body. Also notice that the atrial pulses are all equally spaced as are the ventricular pulses. This is what would be observed when a pacemaker is functioning normally.
FIG. 13 is the same EKG strip as FIG. 12 except in this case an RFID reader has been brought close to the pacemaker and/or its leads in the saline test tank. High frequency (HF) electrical activity is present on the baseline. Although not in any way similar to a physiological signal, in this case, the pacemaker has incorrectly over-sensed and interpreted the RFID signal as a normal heart rhythm and has completely shut off (inhibited) pacing output (atrial and ventricular inhibition). If the patient in whom the pacing device had been implanted had no intrinsic heart electrical rhythm all of the time (was totally pacemaker dependent), or just some of the time (was partially or intermittently pacemaker dependent), the inhibition shown in FIG. 13 either would or could be immediately life-threatening, respectively. Very early on in one of the inventor's experience as a physician with demand pacemakers, this resulted in the death of a very intermittently dependent patient farmer when he was driving his tractor just as he had done many times previously without difficulty. Several identical model demand pacemakers from the same company were then tested while in the shirt pocket of someone sitting in the driver seat of the same model tractor. Inhibition was 100% in all cases caused by EMI from the tractor motor ignition system. It is important to emphasize that intermittent pacemaker dependency (and potentially dependency on other life supporting devices) is common and by its very nature, is under-reported based on the parallel intermittency of follow-up clinic visits. Even if the patient usually has an intrinsic rhythm, dependency for lust a few minutes during an overlapping period of exposure to EMI will be fatal. Similarly, patients lives and well being are frequently at risk because of lack of availability of accurate medical device and clinical data.
FIG. 14 is a pacemaker EKG strip which illustrates another type of Class 1 response involving continual or prolonged ventricular inhibition with individual episodes lasting longer than 3 seconds. As one can see, there is a ventricular stimulus (V1) at approximately 135.5 seconds and then another ventricular stimulus (V2) at 142.2 seconds. Atrial stimulus pulses are shown as A1 through A7 and are undesirably irregular (however, transient atrial inhibition is not considered life threatening), Any transient ventricular inhibition that lasts 3 seconds or more is considered a Class 1 or potentially life threatening response.
FIG. 15 is an EKG tracing which shows high frequency RFID EMI (HF EMI) on the base line tracing, which an ICD has incorrectly interpreted as ventricular fibrillation and delivered a high voltage shock (HV) after 53.4 seconds (if programmed on pacing was totally inhibited). This is defined as a Class 1 response because undesirable high voltage shocks are not only very painful for the patient, but can also result in serious accidents (an inappropriate ICD shock can knock a patient right off his feet).
FIG. 16 is a pacemaker EKG strip example of a Hayes et. al. Class 2 inhibition showing occasional atrial and ventricular output suppression. This is a Class 2 response because of lack of complete atrial inhibition and because the ventricular pulse inhibition is always at least two seconds, but for less than three seconds in duration. It should be noted that no Class 2 responses were found in any of the FDA testing (all responses recorded in the 2008 FDA study were either Class 1 or Class 3).
FIG. 17 is a pacemaker EKG strip of a typical Class 3 response. Through most of this EKG strip, one can see normal atrial (A) and ventricular (V) stimuli. However, at the 49 second point, there is an approximately 33% lengthening of the A-A (A1 -A2) interval, similar to what is seen with occasional T-wave over-sensing at the atrial electrode. This is of no clinical significance and many patients are unaware of a transient slowing of the stimulation rate.
FIG. 18 is a pacemaker EKG strip example of occasional atrial output stimulus inhibition (A) associated with what is most likely atrial triggered ventricular pacing (V). This would be expected with selective detection of EMI on the more sensitive atrial channel versus the less sensitive ventricular sensing circuitry. Atrial sensitivities are almost always adjusted to a low millivolt setting versus the ventricular sensitivities as the electrical signal associated with ventricular contraction is generally 4 times or greater the discharge associated with contraction of the relatively thin atrial wall muscle. Regardless, intermittent loss of atrial ventricular synchrony would rarely be life-threatening for the patient even if it persisted for a long period of time.
FIG. 19 is an EKG strip example of the injection of a CENELEC wave signal, intended to stimulate a normal biologic cardiac electrical signal, into the saline test tank. In engineering terms, the CENELEC signal is intended to represent normal heart electrical activity although the 11 Hz (660 PPM) frequency seen on the base line is more representative of poorly organized atrial or ventricular fibrillation. It is expected that a normally operating pacemaker will be completely inhibited. In other words, what FIG. 19 should look like is the EMI tracing in FIG. 13. The object of complete pacing output inhibition during normal heart rhythms is to avoid delivery of even occasional unnecessary ventricular stimuli (V), as these could result in competition between paced and intrinsic heart action in a patient. This type of interaction is also considered of minor Class 3 clinical significance.
FIG. 20 is an EKG strip example of a pacemaker apparently not responding properly to the CENELEC injection signal, the presence of which is clearly documented on the base line. The expectation was that the delivery of stimuli, by the pacemaker would be completely inhibited as shown in FIG. 13, but in this case as a normal response to a satisfactory patient-generated heart rhythm. Instead, when the RFID reader was brought close to the pulse generator, the sensing circuit classified the supposedly physiologic signal as EMI and automatically switched into the noise reversion (fixed rate) pacing mode, previously outlined as part of the Hayes et al. Type 3 definition. Thus the pulse generator continued to deliver A-V sequential stimuli at regular intervals similar to what was illustrated in FIG. 12, but because of the sensing circuits being by-passed, the atrial (A) and ventricular (V) stimuli in FIG. 20, would be competitive with a patient's underlying heart rhythm. By competitive, we mean the stimuli would fall randomly onto various portions of the intrinsic cardiac action. This is not particularly desirable because while infrequent, application of electrical stimuli while the cardiac tissues are repolarizing (recharging following a muscular contraction) can be arrhythmiagenic. However, this is considered a lesser risk allowing the pacing circuits to be shut off by EMI, when all patients are potentially dependent at one time or another.
FIG. 21 is a bar graph summarizing the FDA 2006 and 2008 pacemaker test data at LF, HF and UHF RFID frequencies. Unfortunately, in 2006, pulse generator responses to the testing were not identified as clinical Type 1, 2, or 3. In other words, trivial and potentially life-threatening responses were lumped together. For example, for 134 KHz (LF) in 2006, 83% of pacemakers tested showed an EMI response. However, the 2008 test stratified the LF data to reveal that 46% of the pacemakers had a Type 1 (life-threatening) response, whereas 32% had a Type 3 response and 22% of the devices were entirely unaffected (these were all unmodulated CW readers). Note: there were no Type 2 responses. A similar range of responses and effects were noted during the testing of the 13.56 MHz (HF) readers. In 2006, a total of 18% of units were affected by the EMI, whereas in 2008, this was refined to show that 7% of devices had a Type 1 response and 4% had a Type 3 response. In 2006, at UHF frequencies, 6% had a response, but this represents only a single pulse generator. It was later determined that this particular pacemaker model did not have a feedthrough capacitor filter. This situation was since rectified through manufacturer re-design. Accordingly, in 2008, none of the pulse generators were affected by any of the readers transmitting 915 MHz signals.
FIG. 22 is a bar graph summarizing the FDA test data that is very similar to FIG. 21 except its for ICDs. It should be noted that 55% of ICDs tested in 2008 at LF showed a Class 1 reaction. This is unfortunate because 134 kHz is an ideal frequency for continuous reader signal emissions to decode a tag embedded deeply inside of body tissue, within the header block, or even inside the housing of an active implantable medical device.
FIG. 23 is a comparison of all of the different types of LF readers tested by the FDA. The reader numbers 1, 2, 3, 4 and 5 correlate with the same RFID equipment code numbers previously described in FIG. 7. As one can see, for RFID reader #1, which was CW, there was no deleterious effect on any of the pacemakers tested. Reader #3, which has a modulation of 11 Hz, effected the greatest number of pulse generators (61% Class 1 and 31% Class 3). This is not particularly surprising if one refers to FIG. 6 and sees that the most sensitive part of both the pacemaker and ICD sensing curves occur at about 11 Hz. The main take-away or summary from FIG. 23 is that all of the modulated low frequency (LF) readers have the potential to cause dangerous pulse generator responses (Class 1) in essentially every case. As previously explained, pacemakers and ICDs really have no practical defense (EMI Filter) to an LF signal that contains modulation within their passband.
FIG. 24 is a bar graph very similar to FIG. 23 except that it compares LF reader FDA test results for ICDs. Again, use of CW reader #1 had no ill effects. However, with reader #3, which has a modulation of 10.5 Hz (which falls right into the ICD sensing passband of FIG. 6), 81% of all responses were Type 1, that is, associated with high clinical risk.
FIG. 25 is a bar graph which is very similar to FIGS. 23 and 24. However, this is a comparison of pacemaker responses at the 13.56 MHz (HF) RFID carrier frequency. Again, reader #8, which is CW, had no effect on any of the pulse generators tested. However, readers #6 and #10 both with 11 Hz modulation resulted in the most detrimental pulse generator responses (13% Type 1 in both cases).
FIG. 26 is a bar graph which illustrates FDA tests using the same HF readers and carrier frequencies as FIG. 25, but for ICDs. Under these test conditions, ICDs tend to be much less susceptible to adverse reader effects than pacemakers. This is likely due to the fact that ICDs are slower to react in providing a high voltage shock as it takes time to charge their high energy internal capacitor, and before pulse delivery a re-interrogation takes place to make sure the dangerous tachyarrhymia is still present.
FIG. 27 summarizes the threshold distances for EMI reactions in the FDA RFID reader testing. The greatest distance at which any reaction was documented was out to 60 cm. This is of great concern compared to the original cell phone work where it was determined that maintaining a transmitter to pulse generator separation greater than 15 cm would be safe. For Type 1 life-threatening reactions, the greatest distances that could lead to an adverse effect was 40 cm with LF readers, and 20 cm for HF readers. For ICDs, the reactions tended to require closer proximity. For any reaction, the threshold distance was 40 cm. For a Type 1 reaction, the RFID reader had to be held within 12.5 cm of the implant in the saline tank. In no case were there any Type 1 reactions for UHF readers. All of these recorded threshold distances are of particular concern for an RFID reader that is designed to directly interrogate an AIMD such as a pacemaker or ICD (to determine the model number, type, serial number, etc. of the implanted device) in these cases, for example in an ambulance or emergency room, the RFID reader would be held as close as 2 cm to the implanted device. The potential for a life threatening Class 1 response is evident.
Similar concerns are present where other types of LF and HF reader applications in which a pacemaker patient or AMD/AIMD may encounter such a reader in the patient's normal environment. For example, keyless entry systems for automobiles generally operate at LF frequencies. The car itself transmits an LF RFID signal which detects the driver/passenger/patient walking up to the automobile where the automobile goes into an active (pinging) mode generating a powerful LF frequency to detect the approach of the driver or passenger who may also possibly be an AIMD patient. When the person nears the car with the car's RFID tag either in his pocket, associated with a wrist watch or other type of container (like a purse or wallet), the car door will automatically unlock (open). Some new automobiles also incorporate a back-up RFID reader system in the driver's seat. For example, in some models, an RFID antenna is also embedded within the driver's side seat back wherein the car tag is reinterrogated to make sure the correct person is inside the car before the ignition will start (this is an anti-high lacking feature). Of course, all of this is of great concern if the particular driver happens to be a pacemaker or ICD patient.
Accordingly, there is a need for an RFID interrogation/communication system having built-in safeguards for protecting sensitive device electronics against RFID-associated electromagnetic interference (EMI). More particularly, an RFID communication system is needed for protecting active medical devices against RFID-associated EMI. Such systems must be able to identify active and passive medical devices through use of RFID technology without causing the AIMD or PMD to malfunction. The present invention fulfills these needs and provides other related advantages.